FISO

In theory yes ie they are approx 76 and 70% of the way to a rise next GW, if those figs are correct and the Threshold exactly doubles for a 2nd Rise in a GW.

How likely is it that Kane drops twice before the deadline?

Joccki_10 wrote: ↑15 Jan 2019, 18:21 How likely is it that Kane drops twice before the deadline?

I think it's likely, he seems bound to drop today and then there's three good days for a mass exodus.

Finisher1 wrote: ↑15 Jan 2019, 19:48

Joccki_10 wrote: ↑15 Jan 2019, 18:21 How likely is it that Kane drops twice before the deadline?

I think it's likely, he seems bound to drop today and then there's three good days for a mass exodus.

Still far away from a second drop. Might keep him as planned and remove TAA instead.

I don’t get how TAA still hasn’t dropped in price.

Am I right in thinking no-one is getting Sterling in (pretty much) despite the DGW coming up? I'm scared he is going to go up in price but doesn't look that way?!

So Kane dropped to £12.4 after not being close to dropping the entire week (including last night!) and TAA still is £5.2? What a load of nonsense.

Joccki_10 wrote: ↑29 Jan 2019, 07:54 So Kane dropped to £12.4 after not being close to dropping the entire week (including last night!) and TAA still is £5.2? What a load of nonsense.

I've been waiting for TAA to drop for about two weeks. Smacks of intervention by the FPL team.

He did switch from Red to Orange back to Red recently, so his threshold may have changed as a result.

I am surprised that Son hasn't risen this weekend.

Chinese Whisperer wrote: ↑03 Feb 2019, 08:24 I am surprised that Son hasn't risen this weekend.

He’d been flagged whilst away on International duty. When I checked yesterday evening he was already 38% towards a rise. Now he’s up to 69%. By the end of today...?

Chinese Whisperer wrote: ↑03 Feb 2019, 08:24 I am surprised that Son hasn't risen this weekend.

you'll have to wait Easter for that, Father

Looks like the fplstatistics site has gone belly up. Annoyingly they had Pereira at 99.2% towards a rise whilst FFF have him at 70% and CTC at 155%.

I'm more likely to trust the first of those. Interested why such a disparity though.

(Edit) looks like fplstats is back now.

Alexander-Arnold was sold by 584.494 managers (actually more but this is the net difference) in the space of four GWs after GW22. No price fall.

Alexander-Arnold was brought in by 92.046 managers (again a net difference) in the space of two GWs after a 13-minute cameo against Bournemouth. A price rise to £5.3 after a no-show against United.

Is there really nothing FPL can do about their flagging system because this just doesn’t look right at all?!

And again I’ve been punished for waiting too long to make a transfer decision. Aubameyang wasn’t even close to a drop and I’ve now lost £0.4 on him. Cannot trust FPLStats anymore.

Joccki_10 wrote: ↑02 Mar 2019, 06:42 Cannot trust FPLStats anymore.

They are not always accurate but I find the data in FFFix as good as any and easy to read. Not a fan of the FPLstats format and do not mind having to give my FPL password to fix as I can not even remember it myself
Price changes have been easy ( ish) to read this year because of the hive mentality and short sightedness of the masses that seems prevalent this season. I am a cr@p manager but taking my transfers early on hunches and best guess has improved my team and squad value this season. I still cannot pick a good captain ( or even the right players) but I am not overly bothered about value after the blanks and doubles.

Joccki_10 wrote: And again I’ve been punished for waiting too long to make a transfer decision. Aubameyang wasn’t even close to a drop and I’ve now lost £0.4 on him. Cannot trust FPLStats anymore.

PEA's price has been like a boomerang this season, in contrast with 'no-drop' TAA (from which I benefited though I never expected to). There is no doubt that the system is broken. I'm not interested in anyone telling me that such-and-such a flag came on or off or that wildcards were in play, or other such BS. It is broken because it is delivering dumbass results.

We are going to get another TAA situation with Kane, who lost 60% of his ownership (from 2m to 800k) whilst only falling from 12.6m to 12.4m. If Spurs get their act together we could easily see his price rocket now. A similar thing happened with Aguero: his price dropped only 0.3m when his ownership dropped from 3.2m to 1.3m, with the result that an ownership rise to 2.4m (0.8m less than the previous peak) has seen the price rise to 0.3m more than when his ownership was 800k higher. It's just silly. But managers who tend to make 'popular' (read 'herd') transfers will defend the system because it always helps them. It is those that do something different that are more likely to be punished.

Is that really what FPL want? Because that's what their system does.

Joccki_10 wrote: ↑02 Mar 2019, 06:42 And again I’ve been punished for waiting too long to make a transfer decision. Aubameyang wasn’t even close to a drop and I’ve now lost £0.4 on him. Cannot trust FPLStats anymore.

The problem is that if there is some kind of 'manual interference' and the examples Ruth cites suggest something like that, then no algorithm is going to pick that up in its forecasts. To my mind FPLStats is still more accurate than FFFix (although to be fair I've barely looked at the latter since Christmas), although I allow a decent margin for error especially when a player is losing owners rapidly.

The current longstanding system is explicitly designed to make it more difficult for well-owned players to fall too much in price, but relatively easy for them to rise. It's also worth remembering that this mechanic will have been chosen to best suit the masses (in order to increase the popularity of the game) and not to appease hardcore players. It was all much more volatile 7/8 years ago!

The TAA situation is the only one that appears to stand out as different to the norm this year, but I haven't followed the details closely enough to try to be able to come up with a way of defending that one. Could also have just been human error.

I don't think that an injured player tumbling in price and so forcing managers (that have already been unlucky) to sell is necessarily a desirable outcome, so in that sense I don't see that the TAA situation is too much of a black mark in the context of FPL's aims to make the game accessible.

Ruth's figures of 60% drops in Ownership causing only .2 and .3m drops seems odd or a hell of a lot of NTO's to be lost on the night of each drop.

However, popular players have always risen quicker then they fell as the percentage for a drop has always been bigger than the absolute figure for a rise?

Was this morning the first price rises this year?

None seem to be captured by fplstatistics. Let the tweaking of algorithms begin

PUBLIC SERVICE ANNOUNCEMENT!

FPL Towers have changed their price change algorithm. Martial was about 60% in predictor sites but he rose anyway. So for some reason FPL Towers have decided to implement a new algorithm. Or they are just messing around with their manual interventions.

So price changes are now totally random until predictor sites figure out the new algorithm. Of course if FPL Towers are just messing around with their manual interventions, then we have to accept that price changes remain random for the whole season.

FPL Towers have really messed up a season start, haven't they? All those glitches and unnecessary new algorithms.

New season new game! I am all for them shaking things up!

Yeah keeps everyone on there toes

Sent from my H8216 using Tapatalk

Not necessarily a change to the system.

Stemania wrote: ↑14 Aug 2019, 10:40 Not necessarily a change to the system.

Guessing when price changes might happen should not be a part of this game in the first place.

The website only references transfer market activity as a reason for changing prices:

Once the season has started, player prices may change by £0.1m a day, based on activity in the transfer market.

Both the formula used to calculate this change and the time of the change include variable factors, and won’t be revealed to game players.

“and the time of the change” is interesting, Do we think it’s always been worded like that?

The back-end for Crack the Code which was not working properly last season has today been updated using a new statistics machine-learning based model.

Fingers crossed that it will start predicting more accurately. Feedback appreciated and we can look at trying to improve the model once we have some results to work on.

admin wrote: ↑16 Aug 2019, 16:38 The back-end for Crack the Code which was not working properly last season has today been updated using a new statistics machine-learning based model.

By the way, a value of 200 means that our overall system is 100% confident that the player’s price will increase, while a value of -200 means that it’s 100% confident that the player’s value will fall.

The closer the Target is to 0, the higher the confidence of our system that the player’s price will not change.

Generally, any value over 100 should be regarded as a price increase prediction and any value below -100 should be regarded as a price decrease prediction.

RomynPG wrote: ↑16 Aug 2019, 17:36

admin wrote: ↑16 Aug 2019, 16:38 The back-end for Crack the Code which was not working properly last season has today been updated using a new statistics machine-learning based model.

Yes, very....